In related art systems, an advertisement system may learn the interests of an individual user and generate customized advertisements for the user based on the user's interests. For example, related art systems may collect information relating to a user's browsing history, prior viewing history, subscriptions to media programs and/or social media history. The advertisement system may identify that the user has ‘liked’ a photograph of organic skincare products on social media. The advertisement system may then insert advertisements related to beauty products in a media program. The advertisement system may insert an advertisement during a commercial break when the user is watching the media program. However, by relying on a large amount of data relating to the user's Internet activities to identify user interests, the burden for the advertisement system to collect and process this large amount of data may be significant. In addition, for some advertisements that are embedded in a media program, e.g., product placement advertisements, it is unrealistic to customize such advertisements to target an individual user, as the advertised product is usually shown as part of a scene in the media program.